Adaptive Histogram Equalization for Video Contrast Enhancement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing contrast enhancement techniques for video processing are unsuitable for ASIC implementation due to computational complexity and storage requirements, and are prone to artifacts like contouring and saturation, especially when dealing with sharp peaks in the luminance histogram.
Innovation Solution
An adaptive histogram equalization-based approach that divides the histogram into three regions, uses contrast-limited histogram equalization, and inserts guard bands to ensure smooth enhancement, while clamping sharp peaks and redistributing rejected points to maintain histogram consistency, allowing for independent processing of each region with adjustable weighting factors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If histogram equalization is applied to enhance contrast, then picture quality improves, but sharp peaks in the histogram cause artifacts like contouring and saturation
Solution Approach 1:
The histogram is divided into multiple regions (e.g., dark, mid-tone, bright regions) and processed independently using separate CDFs. This segmentation allows each region to be enhanced without allowing sharp peaks in one region to propagate and cause artifacts in other regions, thus improving contrast while reducing contouring and saturation artifacts.
Solution Approach 2:
Different regions of the histogram are processed with different CDFs tailored to their specific characteristics. Each region receives localized contrast enhancement that is adapted to its luminance range, preventing uniform over-enhancement that would cause saturation and contouring artifacts while maintaining local contrast quality.
2Illumination intensity
If complex image processing techniques are used for contrast enhancement, then picture quality improves, but computational complexity and storage requirements increase
Solution Approach 1:
By segmenting the histogram into regions and pre-computing CDFs for each region, the system reduces the computational burden during video processing. The pre-computed CDFs can be stored in lookup tables, significantly reducing real-time computational complexity and storage requirements compared to applying complex algorithms to every pixel.
Solution Approach 2:
The CDFs for each histogram region are pre-computed and stored before actual video processing occurs. This preliminary action allows the system to use simple lookup and interpolation operations during real-time video processing, dramatically reducing computational complexity while maintaining high picture quality enhancement.
Data Source
AI summary
An adaptive histogram equalization-based approach improves contrast in a video signal. For each video frame, the histogram of the pixel luminance values is calculated. The calculated histogram is divided into three programmably-sized regions that are equalized independently of each other. The equalization is performed in a controlled fashion by clamping the peaks of the histogram thereby ensuring limited stretching of sharp peaks. The equalized values are averaged with the original pixel values with a weighting factor that is different for the three regions chosen such that the darker regions are enhanced more than the brighter ones. To ensure smooth enhancement, programmable guard band regions can be defined between the three divisions of the histogram. The statistics calculated from one frame may be used to enhance the next frame to eliminate the need for frame buffers. Many of the calculations may be performed in the inactive time between two frames.